[
https://issues.apache.org/jira/browse/MAPREDUCE-5505?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13769934#comment-13769934
]
Vinod Kumar Vavilapalli commented on MAPREDUCE-5505:
----------------------------------------------------
Ideally we should have the state transition itself to go to SUCCEEDED after
unregister. Let's see if that is very difficult to do. Clearly we'll need more
signalling.
OTOH, I think we should hard-code everywhere to say AM that runs only one job.
Part of the problem is that AM at the top level thinks it can run multiple
jobs, though in practice we only run on job. This could simplify some of the
things in the MR AM.
> Clients should be notified job finished only after job successfully
> unregistered
> ---------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5505
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5505
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Jian He
> Assignee: Zhijie Shen
>
> This is to make sure user is notified job finished after job is really done.
> This does increase client latency but can reduce some races during unregister
> like YARN-540
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira